PDA

View Full Version : Field Control System


Nibbles
06-13-2008, 08:28 PM
I wanted to collect ideas on how you think the field control system might work this year. In the past the lights, team numbers, and operator interface were connected to a system that managed the robot, and that was connected to a central computer that kept scores and judges data, which the front screen was also connected to.

With the change to Wi-Fi, I would like to see high- and low-level outlines of how all this might work this year. Some problems to consider are how robots will find their operator controller over IP, how the Wi-Fi setup will work to prevent interference and prevent anything then a competition robot from connecting, how autonomous and disable will work, and how teams will be able to do the same thing.

I wanted a 5Ghz WiFi network, with a single, FIRST-managed access point that teams plug into via Ethernet for the operator controller, robots connect through a WiFi-Ethernet bridge, which would probable have to be custom manufactured, that would connect to the AP, and DHCP basic information about the arena, maybe using an additional service location server, RFC 887 (http://tools.ietf.org/html/rfc887) specified with DHCP option 11. The robot, and team controller, would then have to discover the field controller server, which would collect team numbers and bring the robot controller and driver controller together.

That is enough of me talking, what do you all want?

EricH
06-13-2008, 09:14 PM
what do you all want?Doesn't matter how it does it, I think we all want a field control system that works the first time, every time, and has no breakdowns, no updates, and no cases of Murphy.

Now, whether we get that is something we'll have to find out in March, unless they test it at some offseasons.

cooker52
06-13-2008, 10:33 PM
Something fast, easy to use, CHEAP, CHEAP, CHEAP, and something you can wire literally anything you want into.

I also want it small (I like small :D )

For the most part, I don't really mind if it uses Wifi or is RC, just as long as it can be wirelessly programmed (so I don't have to play the programming cable game again :ahh: ).

Oh, and if they would give us lighter electrical supplies, that would be great.

Besides that, nothing much :rolleyes:

Nibbles
06-16-2008, 02:30 AM
Doesn't matter how it does it, I think we all want a field control system that works the first time, every time, and has no breakdowns, no updates, and no cases of Murphy.

Now, whether we get that is something we'll have to find out in March, unless they test it at some offseasons.

Ethernet is cheap, so are 10Mbps hubs (switches or anything >=100Mbps will cost you). WiFi should have dropped price, but the industry says customers don't care about price, they want speed! (actually it has a little). CAT5 cable for 10Mbps connections are cheap too (not to be confused with CAT6).

I think the best way to ensure that it doesn't break down is to use the same control system protocol that teams will. The biggest problem I see is the robot not finding the field or controller, as long as a team can verify they have found each other, that should solve most all of the field problems, so long as the field doesn't change modes (e.g. Team controllers discover the robot with one way, the field discovers it a second way).

Even better, make the entire control system free software (free as in freedom) so teams can ensure their robots work with the system, or hold pre-ship or off season competitions. Worst case, we have to passively sniff the field communications during the match and reverse engineer a control system.

thefro526
06-16-2008, 05:51 PM
I'd like to see something that's simple to use and easy to replicate for experimental purposes. I think if the field controller is kept simple initially and then modified over time we'll have as few field failures as possible.

jhersh
06-17-2008, 12:39 AM
I wanted to collect ideas on how you think the field control system might work this year. In the past the lights, team numbers, and operator interface were connected to a system that managed the robot, and that was connected to a central computer that kept scores and judges data, which the front screen was also connected to.

With the change to Wi-Fi, I would like to see high- and low-level outlines of how all this might work this year. Some problems to consider are how robots will find their operator controller over IP, how the Wi-Fi setup will work to prevent interference and prevent anything then a competition robot from connecting, how autonomous and disable will work, and how teams will be able to do the same thing.

I wanted a 5Ghz WiFi network, with a single, FIRST-managed access point that teams plug into via Ethernet for the operator controller, robots connect through a WiFi-Ethernet bridge, which would probable have to be custom manufactured, that would connect to the AP, and DHCP basic information about the arena, maybe using an additional service location server, RFC 887 (http://tools.ietf.org/html/rfc887) specified with DHCP option 11. The robot, and team controller, would then have to discover the field controller server, which would collect team numbers and bring the robot controller and driver controller together.

That is enough of me talking, what do you all want?

I think using DHCP to at least configure the IP addresses would make life easier. Doing something that isn't supported in most off the shelf DHCP servers and clients may not be a good idea, though. That's just one more piece of custom software without enough testing that won't quite work right. ;)

It would also be nice if the driver station could be a DHCP server for the laptop. This way the laptop can remain a DHCP client like it normaly would be to plug into a campus network or something, and can also plug into the driver station to be a dashboard without having to reconfigure the network settings.

Would everyone be OK with needing to set static IP addresses on each piece of equipment on the network? Is that easy enough to do that it's not a problem and not worth developing a system that uses DHCP?

-Joe

Jon Jack
06-17-2008, 01:25 AM
Doesn't matter how it does it, I think we all want a field control system that works the first time, every time, and has no breakdowns, no updates, and no cases of Murphy.

Now, whether we get that is something we'll have to find out in March, unless they test it at some offseasons.

Eric is right. Seems like every FIRST event I've been to has had some kind of field control problem. Field controller problems delay events.

Nibbles
06-19-2008, 01:47 AM
One thing I remembered is how will the robot know which alliance it is on? More important to me is, how will the audience know? The flags really, really, badly bug me (not to mention fly 30 ft every so often when bent just the right way). Something I was thinking of was a two color, cold cathode tube almost identical to those used in the green field lights. These are already sold in switchable blue/UV varieties, with two lights inside a single plastic one. You would hook up a diode, so you just need to use a spike controller or something similar to flip the current and change the active light from red to blue. As an aside, if don't cut or extend the wires off of the cold cathode lights, the wires are insulated, and modifying them will result in cross talk (it must not be plain old DC current that comes out of the transformer) that kills the light.